Influence Maximization in Social Networks Under Seeding Budget Constraints: A Novel Shuffled Flamingo Search-Based Metaheuristic Approach for Viral Marketing

Influence Maximization (IM) in online social networks is a well-known complex optimization problem, having significant applications in viral marketing and E-commerce. This paper proposes a novel meta-heuristic driven solution methodology inspired from the shuffled searching behavior of Flamingos for the IM problem under “K” seeding budget constraint. The proposed approach intelligently amalgamates varied structural heuristics of networks such as communities, closeness and betweenness during iterations of optimization to heuristically prune the search space and accelerate convergence. The expected diffusion value function is also incorporated in our proposed approach using the Independent Cascade (IC) model to quantify reliable estimates of influence spreads. A detailed experimentation is conducted for constrained parameter “K” (seed set size) ranging seed set values in the range 10–50. An exhaustive comparative analysis of the results generated by our proposed approach with seven state-of-the-art IM approaches on five real-world social networks in the size ranges small (few hundred) to large (in thousands) is performed. The comparative analysis provides strong evidence for the significantly improved magnitude of spread estimates by the seeds generated by our proposed approach in contrast to many state-of-the-art IM solution methodologies proving its efficacy.